Struct AsyncComputeQueue 

pub struct AsyncComputeQueue { /* private fields */ }

Vulkan async compute queue manager.

Implementations

impl AsyncComputeQueue

pub fn new() -> Self

Create a new async compute queue manager.

pub fn is_available(&self) -> bool

Check if async compute is available.

pub fn submit_compute(&self, _commands: &[u8]) -> GpuResult<()>

Submit a recorded command payload to the async compute queue.

Note

This crate executes all GPU work through wgpu, which exposes a single unified queue and does not surface the separate async-compute / graphics / transfer queue families that raw Vulkan does. The &[u8] command payload accepted here also has no representation in wgpu’s command-buffer model, so genuine asynchronous multi-queue submission cannot be performed on this backend. Rather than silently return Ok(()) and let callers believe GPU work was dispatched, this reports an explicit error. Use crate::GpuContext / crate::ComputePipeline for real execution.

Errors

Always returns GpuError::unsupported_operation: either no compute queue family was detected, or async multi-queue submission is not implemented on the wgpu backend.

pub fn submit_graphics(&self, _commands: &[u8]) -> GpuResult<()>

Submit a recorded command payload to the async graphics queue.

Note

See AsyncComputeQueue::submit_compute — the same wgpu single-queue limitation applies; this never silently succeeds.

Errors

Always returns GpuError::unsupported_operation.

pub fn submit_transfer(&self, _commands: &[u8]) -> GpuResult<()>

Submit a recorded command payload to the async transfer queue.

Note

See AsyncComputeQueue::submit_compute — the same wgpu single-queue limitation applies; this never silently succeeds. When no dedicated transfer queue is present it defers to the graphics queue, which also reports the unsupported-operation error.

Errors

Always returns GpuError::unsupported_operation.

Trait Implementations

impl Default for AsyncComputeQueue

fn default() -> Self

Returns the “default value” for a type.